Summary 2018-05-29T10:52:28+00:00

Are You Ready to Close the Predictive Gap?

The new generation of OLPP-powered predictive applications are here to stay. They will learn from the past to predict the future and plan optimal business outcomes. These applications will optimize supply chains, optimize preventive maintenance, detect fraud, advise medical professionals, power the Internet of Things, and optimize customer interactions.

These applications are operating on an unparalleled scale of data and therefore require a new architecture and platform. The platform needs to be able to ingest at high velocity, compute analytics in real-time at grand scale and be able to handle needle-in-the-haystack type of queries seamlessly. Moreover it needs to be able to handle simultaneous users concurrently and be resilient in the context of error.

The new Splice Machine OLPP platform delivers on this promise. It is a SQL RDBMS built on a Lambda Architecture. Developers and IT can build new predictive applications easily and cost-effectively on this platform. They can migrate old applications as well to the platform and add new predictive components. And finally they can use the platform to just offload old RDBMS’s and Data Warehouses to benefit from scale-out technology without requiring retraining.

OLPP = OLTP + OLAP + ML + Streaming + Notebooks.

With Splice Machine’s OLPP:

  • Data engineers can make their SLAs to deliver data for reporting and for data scientists to use for Machine Learning.
  • Data scientists get a robust data platform where they can leverage their Spark skills within the ease of a notebook but with a more robust database technology for durable storage and for transformation pipelines. They get built in ML or easily connect to external libraries like TensorFlow or MXNet.
  • Business analysts can use their traditional BI tools such as Tableau, MicroStrategy or Qlik because Splice Machine speaks full ANSI SQL.
  • Application developers no longer have to glue multiple engines together to build predictive applications. They can use their programming language of choice and be able to use all the power of traditional SQL including transactional COMMIT/ROLLBACK. They can use either row-based storage in regular tables for needle-in-the-haystack fast queries or columnar storage for hard core analysis. They can deploy on-premise or just connect to an elastic service in the cloud.
  • Business users have predictive applications that advise them on the most recent data with frequently updated machine learning models that make better decisions.
Download PDF